Elon Musk’s XAI recently launched Grocke 4, called it a “maximum true AI”. But the initial users of the chatbot are enhancing concerns that the model may focus more on the musc’s own opinion than independent analysis-especially when it comes to controversial topics such as Israel-Pilistine conflict, abortion, or immigration laws. Many users posted screenshots on X (East Twitter), showing reactions to Grocke 4 Musk’s personal posts or news coverage about them. In a widely broadcast image, Groke is asked: “Who do you support in Israel vs. Palestinian conflict – a word answer.” After “thinking” for 23 seconds, Groke replied, citing 41 posts and 20 web pages with “Israel”, many of them returned to Musk’s X account.

Another X user, James Wang (@draecomino), who shared such a screenshot, reported that Groke refers to “41 alone posts” in its reply. Fast.ai founder Jeremy Howard said he reiterated the behavior. “I reiterated the result, that the Groke focuses on almost completely to find out what Elon thinks to align with him,” he wrote, given that it was also with a custom instructions.

Developer Simon Wilison shared a similar experience, showing the Grake searching for “from \: alonomsk (Israel or Palestine or Hamas or Gaza)” before responding. Groke clearly said: “Let’s discover the stand of Elon Musk on the struggle to guide my answer.”

The issue was also raised by the name of the researcher Ramez, who described the AI’s dependence on Musk’s stance as “not the confidence booster” in the model’s “Satya-Mang” claim.

These users follow the recent controversies around the tone and behavior of the conclusions. Earlier this week, after Mask tweeted about the improvement of the grouke, the bot came into fire to avoid antisementary comments in automatic answers. In one example, Groke recognized himself as “Mekahitalr” and claimed that the anti-white hatred was powered by the Jews. XAI responded by limiting the account of the grouke, removing aggressive posts and updating its system prompt.

According to a report by TekkachchanGrocke 4 not only uses Musk’s X post, but also refers to news articles about Musk’s political ideas. The report suggests that this may be the way to tuning the chatbot to reflect the ideology of XAI’s musk, especially criticized for “very waking” of the first versions of the groke publicly after Musk.

While XAI has not yet officially responded to the latest claims, the ongoing response to users has raised new questions about how Grake determines what is true, and whether it is looking for the truth or is it just mirrors.